Assessment and Validation

A selected test from experiments carried out by Gutierrez-Montes et al. [26] was used in the validation of FDS (version 6.5.3) to predict the behavior of smoke in atria fires. Experiments were conducted in an atrium of the Centro Technologic del Metal, in Murcia, Spain. The atrium's main dimensions were 19.5 m long, 19.5 m wide, and 17.5 m high, as shown in Figure 4.7. The floor was made of concrete and the walls and roof were made of 6-mm thick steel.


Test facility layout and main dimensions [26].

Regarding the mechanical system installed, four exhaust fans were installed on the roof, each with a nominal flow rate of 3.8 m3/s and a diameter of 0.56 m. Eight grilled-type vents were distributed at the lower parts of the walls, each with dimensions of 4.88 m x 2.5 m.

Gutierrez-Montes et al. [26] conducted three series of tests with different heat release rates (HRRs):

  • • Series 1:1.32 MW,
  • • Series 2: 2.28 MW,
  • • Series 3: 2.34 MW.

The tested parameters in these experiments were, among others, smoke temperature, plume temperature, and temperature of the smoke layer. To assess the validity of the numerical solution algorithm and the procedure, only two parameters were examined: plume and exhaust smoke temperature. One hundred and eighty cells each side were chosen for modeling the 1.32-MW fire (Figure 4.8).

Fire Source

A heptane pool fire positioned at the center of the floor level was set as the fire source. The pan diameter was 0.92 m. The peak HRRs for the tests were approximately 1.32 MW. The variation of HRR with time is shown in Figure 4.9. The test duration was 900 seconds.


Central section and top plane layout showing test apparatus used (highlighted).


Heat release variation with time for each of the three test cases as measured by Gutierrez- Montes et al.

Smoke Temperature Measurements

Plume temperatures were measured earlier by Gutierrez-Montes et al. [26] utilizing 3-mm-diameter bare and sheathed type К thermocouples, while surface temperatures were measured by 6-mm-diameter type К thermocouples. Plume Temperature Measurements

The experimental results of plume temperature are compared with the results obtained by FDS 6.5.3 for the test atria. Both sensors at locations 24 and 28 were placed above the fire source to measure plume temperature at different heights, namely, at 4.55 and 12.55 m, respectively. Figures 4.10-4.12 show the measured temporal temperature variation according to the study by Gutierrez-Montes et al. [26] against the predicted temperature distribution throughout time at both locations. Measured and predicted temperature variations were in close agreement with respect to trend, but differ in details. The observed discrepancies were much less than 5%.


FDS geometry used for validation case.


Sensor 24 temperature distribution with time comparison between the work of Gutierrez- Montes et al. and FDS simulation. Smoke Temperature Measurements

The experimental results of smoke temperature were compared with the results obtained by FDS 6.5.3 for the test atria. Sensor 60 located on the roof was used to measure the smoke temperature, as shown in Figure 4.8. Figure 4.13 shows a good agreement between the measured and the modeled smoke temperature values.


Sensor 28 temperature distribution with time comparison between the work of Gutierrez- Montes et al. and FDS simulation.


Sensor 60 temperature distribution with time comparison between the work of Gutierrez- Montes et al. and FDS simulations.

The variations of plume and smoke temperature were compared with the experimental results reported by Gutierrez-Montes et al. [26]. It was noted that all simulations could adequately predict the trend of plume and smoke temperature.

Computational Results for Design Options


An underground car park, 100 m long, 34 m wide, and 3 m high, was considered in the following design analyses. This geometry was chosen as a module for smoke compartmentation to be investigated and then used as a reference for further future design studies. The car park contained one car park inlet 7m wide, 3 m high on the southwest side of the building for car entrance. Two car park outlets 8 m wide, 3 m high on the southeast sides of the building were available for car exit, as shown in Figure 4.14.

Boundary Conditions and Input Data

The Mesh

The model applies 1,285,000 uniform cubic grid cells. The overall volume of the simulated net car park was 10,200 m3. The mesh domain was divided into 16 meshes to save computational time by performing FDS_MPI. The walls, columns, obstructions, and jet fan shrouds were defined as inert surfaces that were nonreacting solid boundary fixed at 40°C. The dimensions of the computational domain were chosen so that the optimum solution recommended by the FDS (version 6.5.3) manual [7] and as described in Table 4.2 could be achieved. Figure 4.15 shows the recommended computational domain.


Car park used in current work: plan view.


Simulation Mesh Parameters


Mesh 01-16


12.5 mx 17 m x 3m

Uniform cubic cell size


Average Nx


Average Nv


Average N,





Simulated car park.

Smoke Management System

Make-up air: Fresh air in the car park was supplied by a mechanical supply system. Mechanical supply air/make-up air was introduced through six supply fans on the right side of the car park. The supply fans were simulated as three vents with an area of 8 m2 each (4 m wide and 2 m high) and a flow rate of 6.14 m3/s for ducted and impulse ventilation scenarios.

Exhaust system: Two systems were proposed for exhaust system, namely:

  • • In impulse ventilation system along with jet fans, five exhaust fans located at the left side of the car park were used. These exhaust fan stations were simulated as vents with an area of 3 m2 each (3 m wide and 1 m high), and a 17 m3/s flow rate based on an assumption of 30 ACH as per the expected expansion ratio of smoke and after many trials to find the required exhaust capacity to ensure acceptable visibility levels.
  • • In ducted system, 86 exhaust grilles were located at the ceiling level extracting 85m3/s based on 30 ACH assumption based on the expected expansion ratio of smoke. Each grille had a 0.16 m2 free area and a flow rate of 0.5m3/s in case of normal mode and lm3/s in smoke mode.

Jet fans: Eight jet fans were used with 50N thrust force (2m3/s volumetric flow). The upper shroud of the jet fans was located directly below the ceiling. Each fan was simulated with a round cross section of 35cm diameter and a shroud length of 2.9 m. The flow rate was maintained for 180 seconds on the normal ventilation mode at 1 m3/s, and then the smoke mode 2m3/s was used till the end of the simulation.

Car Fire

The design fire curve describes the development of a design fire that can be used in a fire scenario. This curve can be as simple as a constant or as a simple function of time. The design fire curve can also be a complicated sequence of lesser curves for some or all stages of fire development. In this study, a fast-growing T-squared fire (Figure 4.16) was chosen for the car fire as this curve is commonly used in similar car fire studies as in [17].

The effect of sprinklers on the car fire HRR was considered in this study. Consequently, instead of the common 8-MW fire used in sprinklered car parks, an HRR of 4 MW is used as per the British code recommendation. The fire dimensions were 4.4 m long, 1.8 m wide, and 0.5 m high with an HRR per unit area of 505 kw/m2. The car fire was simulated by flaming polyurethane as the burning fuel with a soot yield of 0.1 kg soot/kg fuel. (Tables 4.3 and 4.4).

Sprinkler System

The investigated car park was covered with sprinklers with 4 m distance between every two consecutive sprinklers. The firefighting system used was a wet pipe system. Sprinkler type was standard upright operating between 0.5 and 12 bars. Table 4.5 summarizes the used input for sprinklers modeling based on the current assumptions and simulations.


Relation of t-squared fires to some fire tests.


Fire Growth Constants for Т-Squared Fires



A (Btu/s3)

A (kW/s2)

*8 (S>

Range of tg (s)





tg > 400





150 < tg < 400





tg < 150

Ultra fast






Constant Simulations Parameters



Heat release rate of car fire

4 MW

Fire type

T-squared fire

Fire growth

Fast growing fire


Simulated Sprinklers Parameters

Sprinklers Parameters


Operating pressure

1 bar

Flow rate

80 1/min



Orifice diameter

0.5 inch

Particles diameter


Water particles distribution

Rosin-Rammler lognormal

Particles flux

10,000 per second

Activation temperature


Sprinkler density

0.15 gpm/ft2

< Prev   CONTENTS   Source   Next >